Fanaroff-Riley Classification

The picture: where does the jet light up?

Point a radio telescope at an active galaxy and you do not see a star — you see two enormous structures hanging off either side of the host, sometimes stretching millions of light-years into intergalactic space. They are powered by twin jets of plasma launched from a supermassive black hole. The question Bernard Fanaroff and Julia Riley asked in 1974 was deceptively simple: where along those structures is the radio emission brightest?

For some sources, the glow is brightest near the center, close to the host galaxy, and fades smoothly outward into faint, frayed plumes. Those are FR I — "edge-darkened." For others, the inner regions are relatively faint and the brightest spots sit right at the outer ends, in compact knots called hotspots, beyond which the emission abruptly stops. Those are FR II — "edge-brightened." The whole classification rests on that one observable contrast: does the source glow from the inside out, or from the outside in?

That contrast is not cosmetic. It is a direct fingerprint of how the jet died. An FR I jet has slowed down, gone turbulent, and dumped its energy gradually along the way; an FR II jet has stayed razor-sharp and supersonic until it slammed into the surrounding gas at the far end. The morphology is a fossil record of the jet's fluid dynamics, written in synchrotron light.

The precise definition: the ratio R_FR

Fanaroff and Riley wanted a quantitative, repeatable rule, not an eyeball judgement. They defined a structural ratio using the locations of the two brightest patches on opposite sides of the nucleus:

R_FR = (distance between the two brightest regions)
       ────────────────────────────────────────────
              (total largest angular size)

• FR I — R_FR < 0.5. The brightest regions lie in the inner half of the source: edge-darkened.
• FR II — R_FR > 0.5. The brightest regions lie in the outer half: edge-brightened.

Concretely: if a source spans 200 kpc tip to tip and its two brightest knots are separated by 60 kpc, then R_FR = 60/200 = 0.3 → FR I. If those knots are separated by 180 kpc, R_FR = 0.9 → FR II. The genius of the definition is that it is dimensionless and works at any distance: you do not need to know the source's true physical size to apply it, only the relative geometry on the sky.

The original 1974 sample drew on the 3CR catalogue (the revised Third Cambridge Catalogue of radio sources), observed at 178 MHz. Fanaroff and Riley classified 57 sources and noticed something they had not gone looking for: the morphological type was tightly correlated with radio luminosity.

The luminosity break — the surprise that made it famous

The reason this scheme is one of the most cited results in extragalactic radio astronomy is that the two morphological classes turned out to be split almost cleanly by power. Fanaroff and Riley found that essentially all sources below a 178 MHz luminosity of about

L_178  ≈  2 × 10^25  W Hz^-1 sr^-1   (H0 = 50 km/s/Mpc, 1974 cosmology)
       ≈  ~10^25     W Hz^-1         (modern H0 ≈ 70 km/s/Mpc)

were FR I, and essentially all sources above it were FR II. In magnitude terms this is roughly an absolute radio "luminosity threshold" near 10²⁴–10²⁵ W Hz⁻¹ depending on the measurement frequency (the break shifts modestly when extrapolated from 178 MHz to 1.4 GHz with a typical spectral index α ≈ 0.7, where S_ν ∝ ν^(−α)).

The break is real but not a knife edge. Later work, particularly the Owen-Ledlow diagram (Ledlow & Owen 1996), showed that the FR I/FR II dividing line in the plane of radio power versus host-galaxy optical luminosity is sloped: the break radio power scales roughly as the square of the host optical luminosity, P_break ∝ L_optical². A jet of a given power can stay FR II in a small, gas-poor host but disrupt into an FR I in a giant, gas-rich elliptical. That single observation foreshadowed the modern view that environment, not just intrinsic jet power, sets the morphology.

The physics: collimated shock versus entrained plume

Both FR types start the same way: a relativistic jet of magnetized plasma, launched within a few gravitational radii of the black hole (plausibly by the Blandford-Znajek mechanism extracting spin energy, or Blandford-Payne disk winds), propagating outward at bulk Lorentz factors of order Γ ~ 5–30 on parsec scales. What happens next diverges.

FR II — the supersonic working surface. A powerful jet stays collimated and supersonic relative to both the external medium and its own internal sound speed all the way to the lobe end. There it drives a strong shock — the "working surface" or Mach disk — where the ordered kinetic energy is thermalized and the magnetic field amplified. This is the hotspot: a compact (sub-kpc to few-kpc) region of intense synchrotron emission. Shocked plasma then flows back along the jet to inflate a smooth, edge-defined lobe (a "cocoon"). The advance speed of the hotspot is set by ram-pressure balance between the jet thrust and the ambient gas:

ρ_jet v_jet²  ≈  ρ_ext v_advance²        (momentum-flux balance)

→  v_advance  ≈  v_jet √(ρ_jet / ρ_ext)

Because the jet is light (ρ_jet ≪ ρ_ext), the hotspot advances much slower than the jet flow speed — typically a few percent of c, while the jet itself is relativistic.

FR I — the entraining, decelerating plume. A weaker jet (or one in denser gas) is decelerated by entrainment: it mixes turbulently with surrounding interstellar and intracluster gas, mass-loads, and slows from relativistic to transonic and then subsonic speeds within the inner few to tens of kpc. Once subsonic, it can no longer maintain a sharp working surface; it flares open, becomes buoyant, and dissipates its remaining energy gradually as a turbulent plume. The result is bright emission near the nucleus that fades and frays outward — edge-darkened. The deceleration is often visible directly: in sources like 3C 31 the jet brightens, broadens, and bends within the first ~10 kpc precisely where modelling says it transitions through the transonic regime.

The key numbers

Quantity	FR I	FR II
Morphology	Edge-darkened plumes	Edge-brightened, hotspots + lobes
R_FR	< 0.5	> 0.5
L₁₇₈ (W Hz⁻¹)	≲ 10²⁵	≳ 10²⁵
Total jet power Q (erg/s)	~10⁴³–10⁴⁴	~10⁴⁵–10⁴⁷
Jet speed at large radius	Subsonic / transonic (entrained)	Relativistic, collimated
Typical linear size	10 kpc – ~1 Mpc	50 kpc – several Mpc
Lobe advance speed	n/a (buoyant plume)	~0.01–0.1 c
Accretion mode (usual)	Low-excitation, ADAF-like, λ ≲ 0.01	High-excitation, radiative disk, λ ~ 0.01–1
Host environment	Often rich clusters/groups	Often poorer environments at low z
Prototype	3C 31, M84, M87 (Virgo A)	Cygnus A, 3C 47, 3C 236

A few anchoring figures. Cygnus A, the canonical FR II at redshift z = 0.056 (≈ 240 Mpc), has a total radio luminosity near 10⁴⁵ erg/s, a tip-to-tip size of about 130 kpc, and hotspots only a few kpc across that outshine the rest of the lobes. M87, the canonical FR I in the Virgo Cluster at just 16.4 Mpc, hosts a (6.5 ± 0.7) × 10⁹ M☉ black hole — the one imaged by the Event Horizon Telescope in 2019 — and a kpc-scale optical/radio jet that is bright at the base and fades into the surrounding diffuse halo.

Worked example: classify a source and estimate its jet power

Suppose a survey finds a double radio source at redshift z = 0.1. Its two lobes span 4.0 arcminutes on the sky, and its two brightest knots are separated by 3.6 arcminutes, both lying near the outer ends. Its measured flux density at 1.4 GHz is S = 2.0 Jy.

Step 1 — morphological type. R_FR = 3.6 / 4.0 = 0.9 > 0.5 → FR II (edge-brightened).

Step 2 — physical size. At z = 0.1 the angular-diameter distance is D_A ≈ 400 Mpc, so the angular scale is about 1.85 kpc per arcsecond. A 4.0′ = 240″ source spans roughly 240 × 1.85 ≈ 440 kpc — a large, classic FR II double.

Step 3 — radio luminosity. The luminosity distance is D_L = (1+z)² D_A ≈ 484 Mpc ≈ 1.49 × 10²⁷ cm. The monochromatic luminosity, with a k-correction for spectral index α = 0.7, is

L_ν = 4π D_L² S_ν (1+z)^(α-1)

    = 4π (1.49×10^27 cm)² (2.0×10^-23 erg/s/cm²/Hz) (1.1)^(-0.3)
    ≈ 5.3 × 10^32 erg/s/Hz
    = 5.3 × 10^25 W/Hz   at 1.4 GHz

Extrapolating to 178 MHz (L ∝ ν^(−0.7)) raises this by a factor (1400/178)^0.7 ≈ 4.3, giving L₁₇₈ ≈ 2 × 10²⁶ W Hz⁻¹ — comfortably above the FR break, consistent with the FR II morphology we read off the image. The two diagnostics agree, exactly the self-consistency Fanaroff and Riley found in 1974.

Step 4 — jet power. Using the empirical Willott et al. (1999) scaling between low-frequency radio luminosity and mechanical jet power, Q ≈ 3 × 10³⁸ (L₁₅₁/10²⁸ W Hz⁻¹)^(6/7) W with order-unity (factor-of-a-few) uncertainties, an L of a few × 10²⁶ W Hz⁻¹ implies a time-averaged jet power of order 10⁴⁵–10⁴⁶ erg/s. That is comparable to the entire radiative output of a luminous quasar — but delivered as bulk kinetic energy that heats the surrounding gas rather than as light.

Discovery, people, and the surveys behind it

The scheme is named for two then-graduate-students at the University of Cambridge's Cavendish Laboratory. Bernard Fanaroff (later a leader of South Africa's SKA / MeerKAT effort) and Julia Riley published "The morphology of extragalactic radio sources of high and low luminosity" in Monthly Notices of the Royal Astronomical Society in 1974. They worked with maps from the Cambridge One-Mile and Five-Kilometre telescopes and the 3CR catalogue selected at 178 MHz — the low frequency mattered, because it samples the old, steep-spectrum lobe plasma rather than the flat-spectrum core.

Their result landed at a pivotal moment. Quasars had been identified barely a decade earlier (1963), the cosmic microwave background in 1965, and the physical model of radio lobes as jet-fed reservoirs of synchrotron-emitting plasma was just maturing — Martin Rees and Roger Blandford were developing the twin-jet, beam model of extended radio sources in the early-to-mid 1970s. The FR dichotomy gave that theory a sharp observational target: any successful jet model had to explain why power alone could flip a source between two morphologies.

Key follow-ups: Ledlow & Owen (1996) established the host-luminosity dependence (the Owen-Ledlow diagram); Bicknell (1995) and Laing & Bridle (2002) built the entrainment/deceleration models for FR I jets; and large samples from the VLA FIRST and NVSS surveys, and now LOFAR's LoTSS at 144 MHz and MeerKAT, have extended the classification to hundreds of thousands of sources and revealed intermediate and hybrid classes.

Variants and edge cases: FR 0, HYMORS, and remnants

• FR 0. Compact radio sources with bright cores but little or no extended emission, far more numerous than classical FR Is in the local universe. They may be FR Is whose jets are intrinsically weak, intermittent, or recently switched on — never growing large lobes.
• HYMORS (hybrid morphology radio sources). FR I on one side, FR II on the other. Since both jets share one engine, hybrids are powerful evidence that the asymmetric environment drives the difference: one jet meets denser gas and disrupts while its twin stays collimated.
• Wide-angle tail (WAT) and narrow-angle tail (NAT) sources. FR I sources in clusters whose plumes are bent by ram pressure as the host galaxy moves through the intracluster medium — head-tail morphologies that double as wind socks for cluster gas flows.
• Restarted / double-double radio galaxies (DDRGs). A second pair of inner lobes nested inside an older, fading outer pair, recording episodic jet activity — direct evidence that AGN jets cycle on and off.
• Remnant / dying sources. Lobes that have lost their fuel; the hotspots fade first (no fresh jet thrust), the spectrum steepens from radiative + adiabatic losses, and the source relaxes. LOFAR low-frequency surveys are especially good at finding these aged reservoirs.

Common misconceptions and subtleties

• "FR I and FR II are different kinds of black hole." No. The dominant view is that the same class of supermassive black hole produces both; the morphology is set by jet power, accretion mode, and environment. The same engine can even make a hybrid.
• "The luminosity break is a sharp law." It is a strong trend, not a step function. The Owen-Ledlow diagram shows the break depends on host-galaxy optical luminosity, and the scatter is real. There is no single universal L_break independent of environment.
• "Edge-brightening means the FR II is more luminous everywhere." Edge-brightened refers to the spatial distribution of brightness within the source, not the total. It says the peaks sit at the outer ends — the hotspots — not that every part of an FR II outshines an FR I.
• "Hotspots are where the jet hits a literal wall." They are standing shocks at the jet's working surface against the intergalactic gas, not collisions with a solid object. The "wall" is ram pressure from a tenuous (n ~ 10⁻³–10⁻⁴ cm⁻³) medium.
• "FR I jets are non-relativistic from the start." They are launched relativistically, like FR II jets; they merely decelerate to subsonic speeds within the host through entrainment. The base of an FR I jet still shows relativistic beaming asymmetry.
• "Classification needs the source's true size." R_FR is a ratio of on-sky distances, so it is distance-independent. You only need the relative geometry, which is why the scheme has aged so well across redshift.